The Journal of Organic Chemistry
2146 LINNAND CIGANEK
Tetracyanoethylene Oxide. IV. Nucleophilic Ring Opening W. J. LINNAND E. CIGANEK Contribution No. 1473 from the Central Research Department, E. I . du Pont de N m u r s and Company, Wilmington, Delaware Received June 88,1968 Most reactions of nucleophiles with tetracyanoethylene oxide (TCNEO) involve fragmentation of the ring. Examples are given of reactions with amines, thiocarbonyl compounds, benzylideneaniline, and benzophenone azine. The amine reaction gives the same products as the reaction of the amines with carbonyl cyanide. The products from TCNEO and thiocarbonyl compounds depend on the substituents attached to the thiocarbonyl group. Thus, thiourea gives a stable thiouronium ylide whereas with thiobenzophenone, sulfur is eliminated with the formation of a dicyanoethylene. The reaction of TCNEO with benzylideneaniline takes two pathways. One gives an addition product with loss of hydrogen cyanide. The second involves ring scission to form a 1,3dipole that adds a second molecule of benzylideneaniline. The benzophenone azine reaction, however, t’akes a different course and gives nitrogen, l,l-dicyano-2,2-diphenylethylene,and the epoxide of this olefin. Acetic anhydride and acetyl chloride add to TCNEO to give ring-opened products without fragmentation. Benzaldehyde and TCNEO form a 1 : 1 addition product believed to be a cyclic acetal.
The oxirane ring of tetracyanoethylene oxide (TCNEO) is extremely susceptible to nucleophilic attack. A common result is cleavage into the elements of carbonyl cyanide and dicyanomethylene. Examples of reactions with pyridines and alkyl sulfides have been given.] These reactions may be concerted, i.e., both bonds may break simultaneously. However, if tshe process is stepwise, an initial attack on carbon can be followed by one of two different reaction sequences, either of which could lead to the same ultimate products (Scheme I). The two paths differ in the order in which the C-C and C-0 bonds break. These paths can only be distinguished if it is possible to intercept either A or B. SCHEMEI
+
Y:
/”\
(NC),C--C(CN),
\
path b
path a
I +
[&(CN),C(CN),Y] A
[(NC)2COC(CN)2$l B
/
\ CO(CN)Z
+
+-
YC(CN),
Whereas the previously reported reaction with pyridine derivatives is fairly clean, reaction with simple primary, secondary, and tertiary amines is more complex. Yields of products are generally low and it is often not possible to account for all the molecular fragments. One of the reaction products of TCNEO and aniline is N-(cyanoformy1)aniline (1). This is also obtained from reaction of carbonyl cyanide and aniline,3 and it is likely that the TCNEO reaction also involves the intermediacy of carbonyl cyanide. The complexity of the reaction is indicated by the isolation, in small yields, of the anil of carbonyl cyanide (2). An oxidation must be involved to explain this product. Closely related to this reaction is that of TCNEO with benzaldehyde phenylhydrazone. The anil 2 is also a product of this reaction along with l-cyanoformyl-1-phenyl-2-benzylidenehydrazine (3) and small amounts of benzonitrile and benzaldehyde. The structure of 3 was established by an independent synthesis from the hydrazone with carbonyl cyanide and the anil 2 was prepared from malononitrile and nitrosobenzene. The initial reaction presumably results in t,he scission of TCNEO into carbonyl cyanide and the unstable zwitterion 4. The carbonyl cyanide produced in this first step is undoubtedly the source of 3, whereas the TCNEO
We have now observed additional examples of nucleophilic ring-opening reactions of TCNEO. Although some of these are similar to the earlier reported reactions and involve rupture of two bonds, there are two clear-cut instances that follow path a. There are cases also suggesting that path b ring opening can occur, but these are ambiguous. Negative charge is stabilized in both A and B, and a delicate balance may be swung in either direction depending upon the attacking reagent. Another difficulty is the possibility of nucleophilic attack on oxygen. Brown and Cookson2 isolated an unsymmetrical adduct of TCNEO to the 9,lO positions of anthracene. This was attributed to initial attack on oxygen, but it is not possible to distinguish this from a path a reaction. Reaction with Amines.-Xothing can be inferred about the direction of ring opening of TCNEO by amines. Both a C-C and a C-0 bond are broken. (1) W. J. Linn, 0. 15’. Webster. and R. E. Benson, J . A m . Chem. Soc., 87, 3651 (1965). (2) P. Brown and R. C. Cookson, Tetrahedron, 14, 2551 (1968).
CsHsNH2 -+CsHsNHCOCN 1
TCNEO
C~H&H=NNHC~HS -+2
+ C6H,N=C(CN)g 2
+ CsHsCH=NNCsHs
-t
I
J. CO(CN)z
COCN 3
3
C6HsCN
+ CsHsCHO
breakdown of 4 may be responsible for formation of carbonyl cyanide anil and benzonitrile. The complications in such a process are indicated by the low yields of all the products. + CsH6CH=NNHCeH:, ---+ [CsHsCH=NH] I
-b(CNh 4
+2
J. [OI
(3) R. Malachowski and J. Jankiewicz-Wasowska, Rocnniki Chem., 468 34 (1951).
Vol. 34, No. 7 , July 1969
TETRACYANOETHYLENE O X I D E
Products of the reaction of a tertiary aliphatic amine, e.g., triethylamine, with TCNEO are unstable and decompose on attempted purification. It was expected, however, that from dimethylaniline one might obtain bis ( p - dimethylaminophenyl) dicyanomethane (S), the product of Condensation of the amine and carbonyl ~ y a n i d e . ~This is the case although the yields from TCNEO are small and no products were isolated to account for the remainder of the molecule.
CN
&N 5
Reaction with Acetic Anhydride and Acetyl Chloride. -These reactions afforded the first example of nucleophilic ring opening of TCNEO proceeding via path a, the C-C bond remaining intact. Addition of a catalytic amount, of acetate ion to a solution of TCNEO in acetic anhydride results in the precipitation of the diacetate of tetracyanoethylene glycol ( 6 ) , identified on the basis of elernental analysis and spectra. Similarly, 7 is obtained when a drop of pyridine is added to a suspension of T'CNEO in acetyl chloride.
/
\
(NC),C-C(CN),
CN CN
8
9
2147
10
Middleton has described the reaction of 2,2dicyano-3,3-bis(trifluoromethyl)ethylene oxide with thiocarbonyl compounds.6 Tetracyanoethylene oxide reacts similarly with thiourea, thioacetamide, and N,T\"diphenylthiourea to give compounds 8-10, respectively. Side reactions involving carbonyl cyanide were avoided by using ethanol as a solvent, which reacts with carbonyl cyanide to yield ethyl cyanoformate and hydrogen cyanide. TCNEO also reacts rapidly with thiobenzophenone in boiling benzene giving l ,l-diphenyl2,2-dicyanoethylene (1 1) in 68% yield. The reaction occurs more slowly at room temperature in acetonitrile or in ethanol. In the latter instance 3,3,5,5,-tetraphenyl-l,2,4-trithiolane (12) can be isolated in substanbial yield. This compound has been reported' as a minor product in the autoxidation of thiobenzophenone. At elevated temperatures the trithiolane decomposes into thiobenzophenone and sulfur; thus it was not isolated in the reaction at 80". SCHEME I1 0
0
T'ClC-CO-CCH3 I I II
GHP
I
I
6 N CN 7
The preservation of the carbon-carbon bond of the epoxide ring is probably due to solvation by large amounts of the anhydride or acetyl chloride which can trap the anion as scion as it develops. 0
0
, 11 CHSC t
II
OCCH3
A possible mechanism for these reactions involves an unstable dicyano episulfide 13 (Scheme 11). Reaction between 4,4'-bis(dimethy1amino)thiobenzophenone (14) and TCNEO occurs even at -50" yielding the corresponding dicyanoethylene derivative 15 in 79% yield; elemental sulfur was also isolated.
I
6Ac
Reaction with Unsaturated Nucleophiles.-The ability of TCNEO to add to double and triple bonds4 led to the investigation of molecules containing other types of unsaturation. Some preliminary work of this kind has been reported.6 However, each system behaves differently and there are interesting aspects to each. (4) W. J. Linn and R. IC. Beneon, J . A m . Chem. SOC.,87, 3657 (1965). (5) E. Cigsnek, J . 07s. Chem., S O , 4198 (1965).
15
Reaction with Carbonyl Groups.- Unlike the thiocarbonyl compounds, benzaldehyde reacts with TCNEO only at elevated temperature. A 1 :1 adduct could be isolated in about 10% yield. Because of the reaction conditions, it was first assumed that the mechanism of the reaction was similar to that observed for the addi(6) W. J. Middleton, ibid., 81, 3731 (1966). (7) A. Sohiinberg, 0. Sohutz, and S. Nickel, Chem. Ber., 61,2175 (1928); H. Staudinger and H. Freudenberger, ibid.. 61, 1838 (1928).
2148 LINNAND CIGANEK
The Journal of Organic Chemistrg
tion of TCNEO to olefins.8 Therefore the expected product was the dioxolane 16. However, mild hydrolysis of the adduct with dilute hydrochloric acid gives benzaldehyde in good yield.g It is conceivable that hydrolysis of 16 could give benzaldehyde, although this does not seem probable in view of the mild conditions employed. Therefore the hydrolysis experiment favors the alternative structure 17. Spectral characterization
is based on analysis, spectral evidence, and acid hydrolysis to aniline and benzaldehyde. A reasonable mechanism for the formation of the minor product 19 is via a 1,3-dipolar intermediate 20 which can add another molecule of the Schiff base. Evidence for t,his inter-
/"\
C~H~N-CHC~HS + (NC),C-C(CN), CO(CN),
16
+
+
C6H5NeCHC6H5
[;
CGH~NCHCGHS
I
-C(CN),
17
does not distinguish between 16 and 17. The single proton on the five-membered ring shows a sharp resonance peak at T 3.08. This may be contrasted with the hydrogen in the ethylene acetal of benzaldehyde which has its resonance a t r 4.50.1° Although this downfield shift could be interpreted in favor of structure 16, the chemical evidence that points to 17 is stronger. This reaction, therefore, is evidently a second example of ring opening of TCNEO via path a in which the carbon-oxygen bond is preferentially broken. Reaction with Schiff Bases.-Two crystalline products are isolated from the reaction of TCNEO with benzylideneaniline in benzene at reflux temperature. The major product (25% yield) has the composition of a 1: 1 adduct which has lost one molecule of hydrogen cyanide. It is unlikely that this product arises from addition of benzylideneaniline to a species derived from predissociation of the carbon-carbon bond of the epoxide. This mechanism has been shown to operate in addition of TCNEO to olefins,* but at an appreciable rate only above 100". I n this case, nucleophilic attack on carbon followed by ring closure and loss of HCN could only occur with opening of the carbon-carbon bond of the epoxide ring which would give 18. Opening of the carbon-oxygen bond would give rise to a fivemembered ring that could not eliminate HCN. We cannot, however, rule out a structure (18a) in which the ring has opened unsymmetrically by nucleophilic attack on oxygen. However, the longer wavelength
-C( CN )2 20
mediate is furnished by carrying out the reaction of TCKEO and benzylideneaniline in the presence of dimethyl acetylenedicarboxylate. The only product isolated (in 62% yield) is a colorless, crystalline compound to which structure 21 is assigned on the basis of analytical and spectral data. The product loses hydrogen cyanide readily on heating to form the pyrrole 22 in 72% yield. + CP,H~NCHC~H,
I -C( CN
+
CHJOZCSCCOZCH,
-
)z
21
22
This experiment provides strong evidence for the dipole 20 as an intermediate in the reaction, but it is of interest that the adduct 18 is not formed in the presence of the added acetylenic dipolarophile. The fragmentation reaction which leads to the intermediate 1,3 dipole may be reversible (Scheme 111). When the acetylenic ester is present, the dipolar addition reaction is faster than ring closure. I n the absence of the acetylene, ring closure competes with the addition of benzylideneaniline, a poorer dipolarophile. Another possible explanation of the results could involve addition of dimethyl acetylenedicarboxylate to 23 followed by elimination of carbonyl cyanide to give 21.
C6H5 18
18a
19
absorption of the product (Arnx 332 mp) favors struct,ure 18 which has the ring oxygen in conjugation with the double bond and nitrile group. The alternative structure should have an ultraviolet spectrum similar to that of cis-cinnamonitrile (A,,, 273 mp) or 2-phenyll-cyclohexenecarbonitrile (Amax 251 mp)." Neither of these has a longer wavelength absorption. The second product, found in only 8% yield, is l12,3,4-tetraphenyl5,5-dicyanoimidazolidine(19). The latter assignment (8)W. J. ],inn, J . Am. Chem. Snc.. 87, 3665 (1965). (9) A referee suggested the hydrolysia experiment. (IO) B. A. Arbuzov, Yu. Yu. Samitov, and L. K. Yieldssheva, Izu. Akad. hauk, SSSR Ser. F i z . , 47, 89 (1963) (for the English translation, see BulE. Acad. Sci. USSR,Phys. Ser., 27, 95 (1963)). (11) W.E. Parham, W. N. Moulton, and A. Zuckerbraun, J . Ore. Chem., 21, 72 (1956).
A somewhat related reaction is that of TCNEO and benzophenone azine. The reaction occurs smoothly in boiling benzene. Nitrogen is evolved nearly quantitatively (91%), and two crystalline products are formed.
I.'ol. 34, S o . 7 , July 1969
TETRACYANOETHYLENE OXIDE 2149
One wts identified as l,l-dicyan0-2,2-diphenylethylene (24) by an independent synthesis from benzophenone and malononitrile. The other product was believed to be l,l-diphenyl-2,2-dicyanoethylene oxide (25), but attempts to synthesize it by direct oxidation of the olefin were unsuccessful. This was not unexpected because alkaline epoxidation of 1,1-dicyanoethylenes invariably gives the epoxyamide. l 2 0
This same product, is obtained from reaction of diphenyldiazomethane and TCSEO. One might suspect that 25 was formed from diphenyldiazomethane and carbonyl cyanide in this latter reaction. However, it has been shown in this laboratory that carbonyl cyanide adds to diphenyldiazomethane giving 26 which is thermally stable and does not decompose to 25 and nitrogen on heating.13 The structure of the product 25 was established by elemental analysis and spectra combined
26
with the observation that reaction with triphenylphosphine gives benzophenone and triphenylphosphonium dicyanomethylide (27).
(C&,),P f
/"\
(C,Hj),C--C(CNX
+
(C&)PCO
-k
(C&)&XCN)~ 27
The mechanism of formation of 25 in the reaction of diphenyldiazomethane and TCXEO is unknown. Possibly, the active reactant is benzophenone azine (the decomposition product of diphenyldiazomethane), although the olefin 24 mas not actually identified as the other reaction product. No evidence is a t hand that bears on the mechanism of the azine reaction, but the seven-membered ring 28 is a possible method of combining the fragments in the proper order to give the observed products. 1
7
Experimental Section Reaction of TCNEO and Aniline.--A solution of 1.86 g (0.02 mole) of freshly distilled aniline in 15 ml of ether way cooled in an ice bath, and 1.44 g (0.01 mole) of TCNEO was added all a t once. The reaction mixture was stirred for 1.5 hr, and the resulting yellow solution was evaporated to leave 2.78 g of light orange solid. This product was dissolved in hot benzene and the solution was cooled. The precipitated solid was collected by filtration. The light-colored product, mp 123-126', weighed 0.92 g (6370). The reported melting point of N-(cyanoformy1)aniline is 123-126°.2 The infrared spectrum of the product was identical with that of N-(cyanoformy1)aniline prepared by the reaction of carbonyl cyanide and aniline. Attempts to purify the product by repeated recrystallization led to a gradual lowering of the melting point. Malachowski2 reports neither further purification nor analysis of N-(cyanoformyl)aniline, but does report reaction with ammonia to give phenylurea. Our product also reacts with ammonia to give phenylurea, identical with an authentic sample. In another experiment a solution of 1.31 g (0.0141 mole) of aniline in 5 ml of tetrahydrofuran was added, all at once, to a cold (-30') solution of 2.03 g (0.0141 mole) of TCNEO in 15 ml of tetrahydrofuran. The mixture was stirred a t -20" for 5 min and a t room temperature for 3 hr. The solvent was removed a t room temperature leaving 3.88 g of a pale yellow semisolid. Immediate chromatography of 1.80 g of this product over Florisil gave 0.12 g of yellow crystals (eluted with benzenehexane, 3:2) having an infrared spectrum identical with that of carbonyl cyanide anil (see below). The yield was l27& When this reaction was carried out in the presence of yellow mercuric oxide under otherwise identical conditions, the yield of carbonyl cyanide anil was 150/,. Reaction of TCNEO with Benzaldehyde Pheny1hydrazone.To a boiling solution of 3.60 g (0.025 mole) of TCNEO in 50 ml of benzene was added, over 1.5 hr, a solution of 3.90 g (0.025 mole) of benzaldehyde phenylhydrazone in 50 ml of benzene and 20 ml of tetrahydrofuran. The reaction mixture was then heated a t reflux for 1 hr. The cooled mixture was concentrated (rotary evaporator) leaving a dark semisolid. Chromatography on Florisil (elution with benzene-methylene chloride, 3 : 1) gave 1.01 g (52y0 crude yield) of a partially crystalline yellow compound and 1.97 g (63% crude yield) of a pale yellow solid. The former, on crystallization from cyclohexane, gave bright yellow plates of carbonyl cyanide anil: mp 63-64' (sealed tube); ultraviolet spectrum (in cyclohexane), XmB, 367 mp (e 4900), 281 (5300), and 235 sh (5600); infrared spectrum ICHCla), 2225 (m), 1550 (s), 1160 cm-l (vs). Anal. Calcd for C9HjN3: C, 69.67; €I, 3.27; N, 27.08; mol wt, 155. Found: C, 69.70; H , 3.43; N, 27.28; mol wt, 160 (boiling point elevation in benzene). The second product was recrystallized from benzene-cyclohexane to give pale yellow needles of 1-(cyanoformy1)-1-phenyl2-benzylidenehydrazine: mp 154-154.5'; ultraviolet spectrum (in cyclohexane), Am,, 289 mp (e 21,500); infrared spectrum (KBr), 2240 (w), 1690 cm-1 (s). Anal. Calcd for C1~H1,N30:C, 72.27; H , 4.95. Found: C, 72.27; H, 4.34. The reaction was repeated by heating 1.57 g of TCNEO and 2.00 g of benzaldehyde phenylhydrazone in 8 ml of benzene under reflux for 30 min. The flask was then connected to a cold trap, and the volatile products were distilled a t 0.4-mm pressure, first a t room temperature and then a t elevated temperature. The distillate was transferred to a volumetric flask and diluted to 10.0 ml with benzene. Gas chromatographic analysis of this solution was carried out using a 307' silicone oil coumn a t 150'. I t was shown to contain about 130 mg of a mixture of benzonitrile and benzaldehyde (ratio 4.9:l). A small amount of this mixture was collected and the components were identified by mass spectroscopy. Carbonyl Cyanide Ani1 from Malononitrile and Nitrosoben-
L
-I
28 (12) G. B. Payne. J . Orgr. Chem., S6, 662 (1961). (13) TVe are indebted to Dr. E. L. Martin for this experiment.
zene.-A mixture of 2.3 g of potassium cyanide, 16 ml of t-butyl alcohol, 4 ml of water, 20 ml of ether, and 20 ml of benzene was cooled to -5O, and a solution of 3.1 g (0.047 mole) of malononitrile and 5.0 g (0.047 mole) of nitrosobenzene in 15 ml of benzene and 15 ml of ether was added. The temperature was kept below -5' during the addition, which took 15 min. Stirring a t -5" was then continued for 5 min, and a 15% aqueous sodium chloride solution (100 ml) was added. The layers were separated, and the aqueous phase was extracted with 15 ml of
2150 LINNAND CIGANEK ether followed by 15 ml of benzene. The combined organic phases were washed with water, 10% potassium carbonate solution, and concentrated sodium chloride solution and dried over MgSO,. Removal of the solvents left 4.84 g of a dark liquid which was chromatographed on 130 g of Florisil. The first 600 ml of n-hexane-benzene (4:l) eluted 1.79 g of carbonyl cyanide ani1 containing 3% nitrosobenzene as shown by gas chromatography on a 30% silicone grease column a t 155'. An additional 0.41 g of an 80:20 mixture of the anil and nitrosobenzene was eluted with the next 300 ml of the same solvent mixture. The total yield of the anil in both fractions was 2.07 g (29%). Crystallization of the first fraction from cyclohexane gave 1.29 g of the pure mil, mp 63-64'. Another 0.44 g was obtained from the mother liquor and second fraction. The total weight of purified product was 1.73 g (24%). The product was identical with the carbonyl cyanide anil obtained from aniline and TCNEO and from benzaldehyde phenylhydrazone and TCNEO as shown by comparison of the infrared spectra and by mixture melting point determinations. 1-Cyanoformyl-1-phenyl-2-benzylidenehydrazine from Benzaldehyde Phenylhydrazone and Carbonyl Cyanide.-To a vigorously stirred solution of 8.50 g (0.043 mole) of benzaldehyde phenylhydrazone in 25 ml of glacial acetic acid, 25 ml of benzene, and 40 ml of tetrahydrofuran was added, over 15 min, a solution of 3.68 g (0.046 mole) of carbonyl cyanide in 20 ml of benzene. The temperature was kept below -3" during the addition. The yellow solution was then stirred at 0' for 20 min; the solvent was removed and the residue was dried under high vacuum over solid potassium hydroxide. Crystallization from 250 ml of ethanol gave 8.93 g (83%) of I-(cyanoformy1)-1phenyl-2-benzylidenehydrazine, mp 153-154', identical with the product obtained from TCNEO and benzaldehyde phenylhydrazone (see above). Anal. Calcd for CI6HI1N3O: C, 72.27; H, 4.45; N, 16.86; mol wt, 249. Found: C, 72.33; H, 4.49; N , 16.71; mol wt, 267. Bis(p-dimethylaminopheny1)dicyanomethane from TCNEO and Dimethylmiline.-A solution of 2.53 g (0.021 mole) of dimethylaniline in 5 ml of glacial acetic acid was cooled in an ice bath. To this solution I .OO g (0.007 mole) of TCNEO was added in small portions. The reaction mixture first turned light green, then a very dark green, and finally reddish brown. After a short time a yellow solid began to precipitate. After 1 hr this yellow solid was collected by filtration and dried to give 0.45 g of crude product. In a similar preparation which had been allowed to stand for only 10 min, the crude product weighed 0.23 g. These two solids were combined, boiled with ethanol, and filtered. The insoluble portion, 0.37 g, was recrystallized from n-butyl alcohol to give light yellow crystals of bis(p-dimethylaminophenyl)dicyanomethane, mp 195-196'. This material was identified by comparison of its infrared spectrum with that of an authentic sample.8 Bis(p-dimethylaminopheny1)dicymomethme from Carbonyl Cyanide and Dimethy1aniline.-The addition of N,N-dimethylaniline to a solution of carbonyl cyanide in acetic acid results in a strongly exothermic reaction and the production of deep yellow solutions from which greenish yellow crystals separate. Recrystallization of the crystals does not remove the color but passage of a methylene chloride solution of the compound over neutral alumina removes a small amount of an intensely yellow impurity. Colorless crystals, mp 192-193', can be obtained from the colorless solution obtained by ~hromatography.1~ Anal. Calcd for C19HzoN4: C, 74.97; H , 6.62; N, 18.41. Found: C, 74.93; H, 6.71; N, 18.35. Tetracyanoethylene Glycol Diacetate.-To a solution of 2.88 g (0.02 mole) of TCNEO and 10 mi of acetic anhydride was added two small drops of pyridine in 1.0 ml of glacial acetic acid. A precipitate began to form after approximately 15 min. The reaction mixture was stirred overnight a t room temperature and then filtered. The solid was washed with water and air dried to give 3.52 g of crude product which was recrystallized from ethylene dichloride to give 3.24 g (66%) of colorless crystals of tetracyanoethylene glycol diacetate, mp 194.5-195'. Anal. Calcd for CloHsN40,: C, 48.78; H, 2.45; N , 22.76. Found: 48.94; H, 2.57; N, 23.15. The proton magnetic resonance spectrum in deuterioacetone shows only a single peak a t 7 7.38. (14) We are indebted to Dr. E. L. Martin for this experiment. Malnohowski' reports that this product melts at 192-193°.
The Journal of Organic Chemistry 2-Chlorotetracyanoethyl Acetate.-A suspension of 2.88 g (0.02 mole) of TCNEO in 15 ml of acetyl chloride was cooled in an ice bath and stirred. One drop of pyridine was added. After approximately 5 min all the TCNEO had gone into solution. The pale yellow reaction mixture was allowed to stand in the ice bath for 1 hr and then poured onto cracked ice to hydrolyze the excess acid chloride. The solid was collected by filtration, air dried, and recrystallized from carbon tetrachloride to give 2.03 g of tan needles, mp 85.540'. Repeated recrystallization gave an analytical sample, mp 90-91 '. Anal. Calcdfor CsH3C1N4O2: C, 43.16; H, 1.81; C1, 15.93. Found: C, 43.48; H, 1.58; C1, 15.68. The proton resonance spectrum in CD,Cl shows only a single peak a t 7 7.54. S-Thiouroniodicyanomethanide (8).-A solution of 5.79 g (0.0402 mole) of TCNEO in 50 ml of ethanol was cooled to - 10'; thiourea (3.05 g, 0.0402 mole) was added, with stirring, the temperature being kept below -5'. The resulting faintly yellow solution was concentrated to dryness (rotary evaporator) and the residue was dried under high vacuum. The yield of crude S-thiouroniodicyanomethanide was 5.51 g (98Yo). The infrared spectrum was identical with that of an authentic sample.& 2-Amino-1,l-dicyanopropene (9).-To a cold stirred solution of 1.91 g (0.0133 mole) of TCNEO in 15 ml of ethanol was added, over a period of 20 min, 1.00 g (0.0133 mole) of thioacetamide. The temperature was kept below 0"; after ca. 10 min, a yellow precipitate formed. The mixture was stirred a t 0' for another 10 rnin and then filtered to give 0.83 g of a yellow solid, and from the filtrate, 0.88 g of an ill-smelling semisolid. Both products had similar infrared spectra. A small sample of the solid was heated in ethanol and the solution was decanted from the rubberlike insoluble material (presumably sulfur). On cooling, 2amino-1,l-dicyanopropene deposited as colorless needles. Its infrared spectrum was identical with an authentic sample prepared by Middleton.6 4-Amino-5-cyano-3-phenyl-2-phenylimino-2,3-dihydrothiazole (10).-To astirred solution of 3.00 g (0.0208 mole) of TCNEO in 50 ml of ethanol was added, over 20 min, 4.75 g (0.0208 mole) of finely powdered N,N'-diphenylthiourea. The temperature was kept below 30' during the addition. The mixture was stirred a t room temperature for 1 hr, and the yellow precipitate was collected by filtration. The yield of 4-amino-5-cyano-3phenyl-2-phenylimino-2,3-dihydrothiazolc was 4.57 g (75%). Its infrared spectrum was identical with that of a sample prepared by Middleton.6 Reaction of TCNEO with Thiobenzophenone in Benzene a t 80'.-The reaction was carried out in a 100-ml three-necked flask fitted with pressure-equalizing dropping funnel and a 15-cm Vigreux column set up for distillation. To a boiling solution of 3.50 g (0.0243 mole) of TCNEO in 50 ml of benzene was added with stirring, over 10 min, a solution of 4.30 g (0.0218 mole) of thiobenzophenone in 50 ml of benzene. Nitrogen was passed through the reaction mixture, and solvent was distilled continuously during and for about 15 min after the addition. A total of 60 ml of a yellow distillate was collected. The reaction mixture was concentrated to dryness (rotary evaporator), the distillate being caught in a solid carbon dioxide-acetone trap. The two distillates were combined and diluted to 100 ml with benzene. Analysis by gas chromatography (30% Triton X-305 on 60-80mesh acid-washed Firebrick a t 52') showed this solution to contain cu. 0.2 g (10%) of carbonyl cyanide (identification by retention time only). The brown semisolid residue, on chromatography over Florisil, gave 3.40 g (68% yield) of 1,l-dicyano2,2-diphenylethylene (elution with benzene), the infrared spectrum of which was identical with that of an authentic sample (see below). A small amount (0.21 g) of 3,3,5,5-tetraphenyl1,2,4-trithiolane was also obtained (see below). Reaction of TCNEO with Thiobenzophenone in Acetonitrile at Room Temperature.-A solution of 2.95 g (0.0149 mole) of thiobenzophenone and 2.50 g (0.0173 mole) of TCNEO in 30 ml of purified acetonitrile was stirred at room temperature under nitrogen for 24 hr. The yellow precipitate which had formed in the dark green solution was collected by filtration. The yield of crude 3,3,5,5-tetraphenyl-1,2,4-trithiolaneso obtained was 1.08 g (51%). It was purified by chromatography over Florisil (elution with benzene-methylene chloride, 7 :3) and crystallization from methylene chloride-petroleum ether (30-60") to give colorless crystals, mp 122-124' dec, lit.' mp 124'. Bnal. Calcd for C1BH20SR:C, 72.85; H, 4.71; S, 22.44. Found: C, 72.72; H, 4.62; S,22.45.
Vol. 34, N o . 7 , July 1969 The filtrate was concent.rated to dryness (rotary evaporator) and part of t,he residue was chromatographed over Florisil, yielding 165 mg (extrapolated yield: 48%) of 1,I-dicyano-2,2diphenylethylene, identified by its infrared spectrum (see below). TCNEO was recovered in 54Y0 yield by sublimation of the other part of the residue. A similar run, using et,hanol instead of acetonitrile as the solvent,, gave 1,l-dicyano-;!,2-diphenylethylenein 47% yield and 3,3,5,5-tetraphenyl-l,2,4.-trithiolane in 53% yield; the latter had to be isolated by chromatography; it did not precipitate from the ethanol solutiort. l,l-Dicyan0-2,2-diphenylethylene.-Dry ammonia was passed into a solution of 54 g of benzophenone (0.31 mole) and 20 g of malononitrile (0.30 mole) in 180 ml of anhydrous ethanol for 1 hr. The mixt,ure was left. standing a t room temperature for 4 days, but no 1,l-dicyano-2,2-diphenylethylenecrystallized as claimed by previous workers.15 The solution was poured into 400 ml of water and the mixture was extracted with three 150-ml portions of benzene; the combined extracts were washed with 10% potassium hydroxide solution, water, and concentrated sodium chloride solution, and dried over magnesium sulfate. Removal of the solvent left 62 g of a pale yellow liquid to which 50 mi of ethanol was added. After standing a t room temperature for 20 hr, the colorless crystals of l,l-dicyano-2,2-diphenylethylene, mp 140-141.5' (lit.'6 mp 136"), were collected by filtration. A further crop was obtained by adding 50 ml of petroleum ether (30-60") to the mother liquor and keeping the mixture in a cold room for 20 hr; total yield: 8.91 g (12%). An analytical sample, mp 141-142", was obtained by two crystallizations from ethanol; ultraviolet spectrum (in cyclohexane), A, 315 mp ( E 15,800), 275 sh (10,500), and 228 (10,400). Anal. Calcd for ClJlloNs: C, 83.46; H, 4.38; N, 12.17. Found: C, 83.97; H, 4.46; N, 12.09. l,l-Dicyano-2,2-bis(4-dimethylaminophenyl)ethylene (15).A mixture of 10.0 g (0.035 mole) of 4,4'-bis(dimethy1amino)thiobenzophenone, 10.0 g of malononitrile, 5.0 g (0.036 mole) of anhydrous potassium carbonate, and 100 ml of n-propyl alcohol was stirred under reflux for 1 hr. The reaction mixture was cooled with ice, 100 ml of water was added, and the precipitate was collected by filtration, washed with cold water and methanol, and dried. The l,l-dicyano-2,2-bis(4-dimethylaminopheny1)ethylene so obtained (10.80 g, 97%) had mp 254257"; it was 97y0 pure ias indicated by its ultraviolet spectrum. Recrystallization from glacial acetic acid (40 ml/g) gave purple crystals of the pure product, mp 259-260", in 82% yield (overall); ult,raviolet spectrum (in acetonitrile), Amax 430 mp ( E 48,550), 320 (4000), 310 (3700), 265 (22,900). Anal. Calcd for CpallsoHc:C, 75.92; H , 6.37; N, 17.71. Found: C, 75.85; H, 6.42; N , 17.68. Reaction of TCNEO with 4,4'-Bis(dimethylamino)thiobenzophenone.-A solution of 0.92 g (0.0064 mole) of TCNEO in 5 ml of tetrahydrofuran and 10 ml of ethanol was cooled to around - SOo, and a soluthn of 1.53 g (0.0054 mole) of 4,4'-bis(dimethy1amino)thiobenzophenone in 30 ml of chloroform was added with vigorous stirring over 25 min. The red color of the thioketone changed t,o green immediat,ely. The mixture was allowed to warm to room temperature (1 hr); it was then stirred a t room bemperabure for 30 min. The solvents were removed leaving 1.95 g of a brown semisolid. A small amount (46 mg) of this solid was dissolved in 50 ml of acetonitrile; the insoluble portion (3.7 mg of a rubber-1ik:e material) contained 88.8% sulfur as shown by microanalysis; total yield of elemental sulfur: 140 mg (8lSz). The acet,onitrilesolution had an ultraviolet spect>rum identical with that OF l,l-dicyano-2,2-bis(4-dimethylaminopheny1)ethylene: the ! i value of the 430-mp band was 120, indicating that the product contained 1.52 g (8970 yield) of the ethylene derivative. Recrystallization of 1.31 g of the crude product from 40 ml of glacial acetic acid gave 0.90 g (79%) of 1,l-dicyano-2,2-bis(4-dirnethylaminophenyl)ethyleneas purple crystals, mp 256-258". The infrared spectrum was identical wit,h that of an autheniic sample prepared as described in the preceding experiment. 4,4,5,5-Tetracyano-2-phenyl-1,3-dioxolane (17).--4 reaction vessel containing 3.00 g (0.0208 mole) of TCNEO in 25 ml of freshly dist,illed benzaldehyde was alternately evacuated and flushed with nitrogen. I t was heated a t 100-110' for 18 hr. Most of the benzaldehyde was evaporated a t reduced pressure, (15) R .
Sohenck and H. Finken,Ann. Chem., 463, 27 (1928).
TETRACYANOETHYLENE OXIDE 2151 and the residual solid was dried on a porous plate. The dried residue was sublimed at 80-90" (0.1 mm) until sublimation ceased. This first fraction was mainly recovered TCNEO as indicated by the infrared spectrum. However, some of the desired adduct waa present. The sublimation was continued at 110'. The second fraction contained essentially no TCNEO. This product was recrystallized once from 2-propanol to give 0.62 g of brown needles which were sublimed a t 190' (0.1 mm) to give 0.53 g (10%) of colorless crystals, mp 140-142'. Further recrystallization from 2-propanol raised the melting point to 141-142'. Anal. Calcd for CIsH~N4Oz:C, 62.40; H, 2.42; N , 22.39; mol wt, 250. Found: C, 62.15; H , 2.48; N, 22.11; mol wt, 238-240. Acid Hydrolysis of Benzaldehyde-TCNEO Adduct.-A suspension of 0.101 g of the 1: 1adduct of TCNEO and benzaldehyde in 5 ml of 57, hydrochloric acid was warmed overnight on a steam bath in a nitrogen atmosphere. The solid disappeared and was replaced by small drops of oil immiscible in the aqueous phase. The mixture smelled strongly of benzaldehyde. The reaction mixture was extracted with methylene chloride and the extract was dried with Drierite. Gas chromatography of the extract on a 6 ft X 0.25 in. column of XE-60 on Gas Chrom RA a t 145' and a helium flow of 100 ml/min showed, besides the methylene chloride peak, a component eluting a t 4.4 min. An authenticsample of benzaldehyde in methylene chloride also eluted a t 4.4 min. Removal of the solvent with a rotary evaporator left a small residue which was dissolved in 6 ml of 50% aqueous ethanol. The solution was heated to boiling for 0.5 min after the addition of two drops of phenylhydrazine. Cooling the solution caused the precipitation of pale yellow needles which weighed 0.050 g (637,) and melted a t 153-155". Recrystallization from 50% aqueous ethanol raised the melting point to 155-156.5'. An authentic sample of benzaldehyde phenylhydrazone had the same melting point and the infrared spectra of the two samples were identical. Reaction of TCNEO and Benzy1ideneaniline.-The reaction was carried out in a 500-ml, three-necked flask fitted with g&inlet tube, two 125-ml dropping funnels, and a 20-cm Vigreux column connected to a 1000-ml receiving flask, the contents of which (45 g of dimethylaniline in 100 ml of glacial acetic acid) was cooled with ice and stirred. To 500 ml of boiling benzene in the three-necked flask were added separately and with stirring, solutions of 25.5 g (0.18 mole) of TCNEO in 75 ml of tetrahgdrofuran and of 46.5 g (0.26 mole) of benzylideneaniline (Eastman White Label, mp 53-54') in 200 ml of benzene. Nitrogen was passed through the reaction mixture, and solvent was distilled out continuously. Addition of the two solutions was complete after 50 min. Benzene (100 ml) was added and solvent distillation was continued for 25 min, when the volume of the dark green reaction mixture was ca. 200 ml. I t was cooled with ice and the yellow precipitate was collected by filtration. It weighed 13.3 g (25%). Crystallization from 85 ml of acetonitrile gave 9.05 g of yellow needles of 2,2,4-tricyano-1,5-diphenyl-A4-oxazoline (18), mp 232.5-234' (sealed tube). Further recrystallization did not increase the melting point. The infrared spectrum (KBr) showed bands a t 2230, 1685, and 955 cm-1; ultraviolet 252 mp (e 21,600) and 332 (6500); spectrum (in acetonitrile), A,, nmr spectrum, two peaks (each split a t least into a doublet) at 7 2.29 and 2.50 (ratio of areas 1:1). Anal. Calcd for C I ~ H I O N ~ O C,: 72.47; H, 3.38; N, 18.78; mol wt, 298. Found: C, 72.62; H, 3.30; N , 18.79; mol wt, 319 (ebullioscopically in benzene). The mother liquor was concentrated, using a rotary evaporator, and to the residue (56.78 g of a black oil) was added, with stirring with a glass rod, 200 ml of boiling ethanol. The cooled mixture was filtered to give 21.16 g of a light tan solid, which was chromatographed on 300 g of Florisil. A pale yellow solid (8.47 g) was eluted with benzene. It was crystallized from acetonitrile, giving 6.48 g (8.5% yield) of 1,2,3,4-tetrapheny1-5,5-imidazolidinedicarbonitrile (19) as almost colorless plates, mp 209-21 1O (sealed tube). The analytical sample melted a t 210-211'; ultraviolet spectrum (in acetonitrile), Xmax 237 mp (e 19,300) and 287 sh (1900); infrared spectrum, weak CN band a t 2240 cm-1 (in KBr); nmr spectrum (in acetonitrile), singlets a t 7 3.92 and 3.34 (area -1 each) and a very complex spectrum from 2.4 to 3.2 (area -20). Anal. Calcd for C W H ~ Z NC, ~ : 81.66; €1, 5.20; N , 13.14; mol wt, 426. Found: C, 81.89; H, 5.29; N, 13.17; mol wt, 451 (ebullioscopically in benzene).
2152 LINKAND CIGANEK The contents of the receiving flask (containing the reaction product of dimethylaniline with carbonyl cyanide) were concentrated to dryness using a rotary evaporator. The residual yellow crystals were heated with 30 ml of ethanol and collected by filtration to give 5.73 g (11%) of bis(p-dimethylaminopheny1)dicyanomethane identified by comparison of its infrared spectrum with that of an authentic sample (see above). Reaction of TCNEO with Benzylideneaniline and Dimethyl Acetylenedicarboxy1ate.-The apparatus used was a lOO-ml, three-necked flask fitted with magnetic stirrer, gas-inlet tube, 100-ml dropping funnel, and a 15-cm glass spiral column set for distillation. To a stirred boiling solution of 3.24 g (0.0225 mole) of TCNEO and 6.40 g (0.045 mole) of dimethyl acetylenedicarboxylate in 50 ml of benzene was added, over 70 min, a solution of 4.10 g (0.0225 mole) of benzylideneaniline in 50 ml of benzene. Nitrogen was passed through the reaction mixture, and solvent was distilled into a receiving flask (containing 8 g of dimethylaniline in 15 ml of glacial acetic acid) a t about the same rate as benzylideneaniline solution was added. Thirty milliliters of benzene was then added, and distillation of solvent was continued for 30 min. The reaction mixture, on cooling with ice, deposited 4.67 g of light tan crystals of dimethyl 2,2-dicyano-1,5-diphenylA~-3,4-pyrrolinedicarboxylate (21), mp 174-176" dec. From the filtrate, on evaporation to dryness and chromatography on Florisil, an additional 0.73 g of the diester was obtained. The analytical sample, prepared by two recrystallizations from benzene, melted a t 178' dec; ultraviolet spectrum (in acetonitrile), Xmax 230 mp ( B 16,000), shoulders a t 265 (2800) and 272 (2500); infrared spectrum (KBr), 2240 (w), 1750, 1730 (s) cm+. Anat. Calcd for CZZH17NIOI: C, 68.21; H , 4.43; N, 10.85; mol wt, 387. Found: C, 68.48; H, 4.56; N, 10.71; mol wt, 429. The yield of bis(p-dimethylaminophenyl)dicyanomethane, isolated from the contents of the receiving flask as described for the reaction of TCNEO with benzylideneaniline, was 1.60 g (2370). No 19 was obtained in the chromatography, nor was any 18 isolated by sublimation of a small sample of the evaporated filtrate (before chromatography) a t 180" and 0.3 mm. Thermal Decomposition of 21.-A sample of 0.61 g of the pyrroline 21 in a 10-ml flask was evacuated to 120 Torr and slowly heated in an oil bath. Smooth evolution of hydrogen cyanide started a t 171' (bath temperature). After 20 min a temperature of 185" had been reached and no gas was being evolved. On cooling, 0.53 g of a colorless solid was obtained. This was recrystallized from 4 ml of benzene-cyclohexane (3: 1) to give 0.40 g (71%) of dimethyl 1,2-diphenyl-5-cyano-3,4pyrroledicarboxylate (22): mp 151-152'; ultraviolet spectrum (in acetonitrile); hmax285 mp ( t lO,OOO), 235 (22,000); infrared spectrum (KBr), 2240, 1740, and 1725 ern-'. Anal. Calcd for C21H16N204: C, 69.99; H, 4.48; N, 7.77; mol wt, 360. Found: C, 70.56; H, 4.43; N, 7.86; mol wt (ebullioscopic in benzene), 370, 371. Reaction of TCNEO with Dipheny1diazomethane.-To a stirred boiling solution of 1.35 g (0.0094 mole) of TCNEO in 40 ml of benzene was added, over 30 min, a solution of 1.81 g (0.0093 mole) of diphenyldiazomethane in 30 ml of benzene. The mixture was then heated under reflux for 10 min. A total of 200 ml (88%) of nitrogen was evolved. In a second experiment, solvent was distilled continuously during the addition; no carbonyl cyanide was detected in the distillate. The solvent was removed a t reduced pressure to give 3.19 g of a dark semisolid, part of which (1.66 g ) was chromatographed over Florisil
The Journal of Organic Chemistry (45 9). In the first 200 ml of benzene, 0.40 g (17%) of 1,1dicyano-2,2-diphenylethyleneoxide was eluted. Two recrystallizations from benzene gave an analytical sample: mp 166-167'; ultraviolet spectrum (in cyclohexane), Xmax 269 mp ( B 685), 262 (850), 258 (750), 256 (755), 226 (11,900); infrared spectrum (KBr), 2245, 1590, 1500, 1240, 1230, 1225,910, and 885 cm-1. Anal. Calcd for ClsHloN~O: C, 78.03; H, 4.10; N, 11.38; mol wt, 246. Found: C, 78.13; H, 4.12; E,11.28; mol wt, 261. An unstable yellow oil was obtained by further elution with benzene and benzene-methylene chloride. Reaction of l,l-Dicyano-2 ,2-diphenylethylene Oxide (25) with Tripheny1phosphine.-A mixture of 0.22 g (0.00093 mole) of l,l-dicyano-2,2-diphenylethyleneoxide, 0.44 g (0.0015 mole) of triphenylphosphine, and 3 ml of 1,2-dichloroethane was heated under reflux for 1 hr. Removal of the solvent (rotary evaporator) gave a dark oil, which was dissolved in 2 ml of hot benzene. Addition of 2 ml of hot cyclohexane and cooling gave 0.20 g of triphenylphosphoniuni dicyanomethylide.16 The mother liquor was concentrated to dryness and the residue was chromatographed over Florisil. Triphenylphosphine (0.19 g ) was eluted with the first 36 ml of benzene-n-hexane ( 1 : l j . Another 60 ml of the same solvent mixture eluted 0.15 g (92% yield) of benzophenone, identified by its infrared spectrum. A small quantity (ca. 0.03 g) of triphenylphosphonium dicyanomethylide (27) was eluted with methylene chloride-tetrahydrofuran (4: 1). The total yield of 27 was 0.23 g (797,). Reaction of TCNEO with Benzophenone Azine .-The reaction was carried out in a 100-ml three-necked flask fitted with a pressure-equalizing dropping funnel and a reflux condenser, the top of which was attached, through a drying titbe, to a 100-ml gas buret. To a boiling solution of 1.6s g (0.0117 mole) of TCNEO in 15 ml of benzene was added, over 8 min, a solution of 1.40 g (0.00389 mole) of benzophenone azine in 20 ml of benzene, and the mixture was then heated iuider reflux until gas evolution ceased (3 hr). A total of 86 ml (at 24', 917, of the theoretical amount) of nitrogen was evolved. The solvent was removed (rotary evaporator), leaving 2.99 g of a tan solid. Chromatography over Florisil gave 0.73 g (767,) of 1,l-dicyano2,2-diphenylethylene oxide (eluted with benzene-hexane, 2 :3) and 0.79 g (88'%) of l,l-dicyano-2,2-diphenylethylene(eluted with benzene). The two products were identified by comparison of their infrared spectra with those of samples prepared previously (see above).
Registry No.-TCXEO, 3189-43-3; 2 , 19769-98-3; 3, 19769-99-4; 6 , 19770-00-4; 7, 19770-01-5; 14, 1226-46-6; 15, 19770-02-6; 17, 19770-03-7; 18, 19770-04-8; 19, 19769-85-8; 21, 19769-86-9; 22, 19769-87-0; 25, 19769-88-1; aniline, 62-53-3; benzaldehyde phenylhydrazone, 585-64-7 ; acetic anhydride, 108-24-7; acetyl chloride, 75-36-5; t hiobenzophenone, 1,450-31-3; benzylideneaniline, 535-51-2; diphenyldiazomethane, 883-40-9; benzophenone azine, 98379-9. Acknowledgment.-We are indebted to Dr. R. E. Benson for many helpful discussions in connection with this work. (16) L. Hoerner and H. Oediger, Chem., Be?., 91, 437 (1958)
